Developmental Psychology 1990, Vol. 26, No. 5, 855-865 Copyright 1990 by the American Psychological Association, Inc. 0012-1649/90/$O0.75 Schematic Knowledge Influences on Memory for Scene Information in ^foung and Older Adults Thomas M. Hess and Susan Jensvold Slaughter North Carolina State University Two experiments were conducted to examine the hypothesis that aging is associated with an in- crease in schema-related influences on memory performance. In both studies, groups of young and older adults studied organized and unorganized visual scenes containing objects that varied in their likelihood of occurrence. Measures of attention allocation and memory for both objects and their relative locations were obtained. Consistent with expectations, attention allocation during study was disrupted more by lack of scene organization in the older adults. In addition, relatively systematic age differences were also obtained for retention, with greater schematic effects observed in the memory performance of older adults than in that of young adults. The results are discussed relative to current views of aging and cognitive change. Most current conceptualizations of adult development em- phasize the multidimensional nature of intelligence. Specifi- cally, it appears that there are at least two distinct groups of abilities, each of which exhibits a different pattern of change over the adult life span. For example, Baltes, Dittmann-Kohli, and Dixon (1984) distinguished between the mechanics and pragmatics of intelligence. They airgued that the basic content- free components of intelligence (i.e., mechanics) become less efficient with age, whereas those relating to specific knowledge systems (i.c, pragmatics) are relatively stable or continue to de- velop. One important implication of this general view of adult development for the study of mnemonic processes is that aging should be associated with an increased reliance on knowledge systems in support of memory functioning as basic memory processes decline in efficiency. Although there has been little systematic investigation of this hypothesis, some evidence does exist that supports this view (e.g, Hess, 1985; Hess, Donley, & Vandermaas, 1989; Waddell & Rogoff, 1981,1987). For exam- ple, Waddell and Rogoff (1981, 1987) showed that age differ- ences in memory are small when test stimuli are meaningfully organized. Our research attempts to provide further evidence relevant to this hyothesis by examining age differences in the effects of knowledge on memory for information contained in visual scenes. Experiment 1 is based on Susan Jensvold Slaughter's master's thesis. We are indebted to Jean Mandler for providing us with copies of her scene stimuli for this experiment, and we would also like to thank Sharon Wallsten for her assistance in data collection and analysis. The research reported in Experiment 2 was supported in part by National Institute on Aging Grant AGO5552-O2 awarded to Thomas M. Hess. We would like to thank Susan Hess, who developed the stim- uli for this study, and Jan Donley, Shirley Floyd, Diane Ursone, and Maureen Vandermaas for their assistance in various phases of data collection. Correspondence concerning this article should be addressed to Thomas M. Hess, Department of Psychology, North Carolina State University, Raleigh, North Carolina 27695-7801. Experiment 1 Work with young adults has shown that scene retention is affected by the relationship between the information depicted in the scene and the world knowledge possessed by the individ- ual. This research has generally been conducted by using a schema-based framework of memory (cf. Mandler, 1984). A primary assumption from within this framework is that knowl- edge structures (i.e, schemas) influence memory processing in a top-down fashion by organizing the perception of a scene and creating expectations about the identity of objects in the scene and the spatial relations between them. Subsequent memory for both types of information is determined by the conformity of the scene contents to expectations. With respect to scene objects, the major impact of schematic knowledge on memory processing appears to be in relation to the liklihood of occurrence of individual objects in specific types of scenes. Research has shown that expected information (e.g, a stove in a kitchen) is attended to less than unexpected information (e.g, a cat in a kitchen), at least initially (Friedman, 1979; Loftus & Mackworth, 1978). This effect is attributed to the fact that objects predicted by the schema activated at per- ception require less analysis for identification than those that are not predicted (e.g, Palmer, 1975). One memorial conse- quence of this differential allocation of attention is that unex- pected information tends to be remembered better (e.g, Fried- man, 1979). This effect appears to be related both to the poorer detail in the representations of expected objects in memory (e.g, Mandler & Parker, 1976; Mandler & Ritchey, 1977) and to the tendency for individuals to supplement these representa- tions with expected but unpresented information (e.g, Fried- man, 1979). Both of these factors would result in poorer dis- crimination memory for expected objects. Knowledge effects on memory for spatial information have been investigated primarily by examining how the disruption of normal spatial relations affects performance. For example, research on memory for objects in organized and unorganized scenes by Mandler and her colleagues (Mandler & Johnson, 855